CN109117993B - Processing method for optimizing vehicle path - Google Patents
Processing method for optimizing vehicle path Download PDFInfo
- Publication number
- CN109117993B CN109117993B CN201810845753.0A CN201810845753A CN109117993B CN 109117993 B CN109117993 B CN 109117993B CN 201810845753 A CN201810845753 A CN 201810845753A CN 109117993 B CN109117993 B CN 109117993B
- Authority
- CN
- China
- Prior art keywords
- vehicle
- path
- time window
- model
- maximum
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
- G06Q10/047—Optimisation of routes or paths, e.g. travelling salesman problem
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- General Physics & Mathematics (AREA)
- Strategic Management (AREA)
- Artificial Intelligence (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Biophysics (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Development Economics (AREA)
- General Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- Biomedical Technology (AREA)
- Health & Medical Sciences (AREA)
- Game Theory and Decision Science (AREA)
- Computational Linguistics (AREA)
- Entrepreneurship & Innovation (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Life Sciences & Earth Sciences (AREA)
Abstract
The invention discloses a processing method for optimizing a vehicle path, which comprises the following steps: a. collecting the order number, the network point distance matrix and the vehicle information; b. setting whether a time window and/or a circuit template is opened or not; c. taking the order number, the mesh point distance matrix, the vehicle information, whether a time window is opened or not and whether a line template is opened or not as constraint conditions, obtaining an optimal path by adopting ant colony genetic fusion algorithm iterative computation, and updating mesh point pheromones; the design reasonably utilizes constraint conditions, and the ant colony genetic fusion algorithm is subjected to replication and transformation after iterative computation to obtain the shortest path, and the result performance index is good.
Description
Technical Field
The invention relates to the field of vehicle logistics, in particular to a processing method for optimizing a vehicle path.
Background
The first study of logistics path planning problem abroad was in the middle page of about 20 th century, and two-figure scientists first proposed a model of the problem of automobile distribution, namely the vrp (vehicle Routing distribution) model. Since the logistics range is getting larger and larger, the practical application value of the problem is getting higher and higher, so that more and more researchers are involved, and scholars in various fields participate. Bodin et al concluded seven major methods of solving these models based on these models. Based on a linear programming idea proposed by G.B.Dantzig in 1947, an implicit enumeration algorithm proposed by B.Jackowski et al in 1985 and an iterative greedy algorithm proposed by J.Culberson in 1992, and other precise algorithms, the small-scale data network point VRP is successfully solved, and the algorithms can accurately solve the optimal solution of the minimum path under the current condition under the condition that the network point number is not large. When the dot size is large, the precise algorithm can not be used any more. In 1995, a particle swarm algorithm was researched by Eberhart and Kennedy according to a cooperative method in the bird group feeding, and an approximate solution is obtained by using a random solution iteration mode. In 1999, DonaldWaters and ronaldh. Eiichi T, Michihikon, Tadashiy and Torul are combined with factors such as traffic and cost in reality to design a double-layer mathematical model so as to optimize the selection of distribution centers. In 2002, Yang-JaJaJang combines Lagrange relaxation algorithm and genetic algorithm to optimize and solve the path planning problem of logistics. In 2003, Jayaraman applied a simulated annealing algorithm to solve the problem of optimizing the path planning of logistics under a model of multiple products, single production base and multiple distribution centers. In 2006, Marie-claude bolduc et al proposed that a heuristic mathematical algorithm be applied to optimize the solution problem under a vehicle routing model with mass production and multiple distribution points with time window constraints, and also be applicable to the vehicle routing problem of multi-level retailers.
At present, the research on VRP under multiple constraint conditions in China is still in the theoretical research and experimental stage. The experimental results can be well expressed under specific conditions, but cannot be used in a large scale, the performance difference of the algorithm is huge according to the change of constraint conditions, the reusability of the algorithm is poor, the optimal solution obtained by singly adopting the ant colony algorithm or singly adopting the genetic algorithm is not accurate, and the performance index is poor through experimental analysis.
Disclosure of Invention
In order to solve the above technical problems, an object of the present invention is to provide a processing method for performing path optimization by iteration using an ant colony genetic fusion algorithm according to the order number, the mesh point distance matrix, vehicle information, whether to open a time window and/or a line template as constraint conditions.
The technical scheme adopted by the invention is as follows:
a vehicle path optimization processing method, comprising the steps of:
a. inputting order number, network point distance matrix and vehicle information;
b. setting whether a time window and/or a circuit template is opened or not;
c. and (3) taking the order number, the mesh point distance matrix, the vehicle information, whether a time window is opened or not and whether a line template is opened or not as constraint conditions, carrying out iterative computation by adopting an ant colony genetic fusion algorithm to obtain an optimal path, and updating mesh point pheromones.
And d, uploading the calculated result to a database of the cloud server.
In the step c, iterative computation is carried out within the range that the iteration upper limit value is not reached, and the method comprises the following steps:
c1, initializing ant colony and ant;
c2, selecting the next network point by the ant according to a deterministic exploration strategy or a stochastic exploration strategy;
c3, judging whether the ants go through all the network points, if yes, entering the step c4, and if not, returning to the step c 2;
c4, obtaining the circulating optimal ants after all the mesh points are processed;
c5, copying and crossing the cycle optimal ants and the global optimal ants obtained in the iterative process;
c6, performing exchange variation and inversion variation on globally optimal ants;
c7, judging whether the global optimal ants are subjected to iterative evolution, if so, entering a step c8, and otherwise, entering a step c 9;
c8, replacing the mesh point information of the original global optimal ant with the optimal solution obtained by iterative evolution, and entering step c 9;
c9, updating the site pheromone;
c10, judging whether the iteration number exceeds the iteration upper limit value, if so, outputting the result, and if not, returning to the step c 2.
And c, establishing a single-vehicle multi-path model or a multi-vehicle multi-path system model or a mixed path model combining the single-vehicle multi-path model and the multi-vehicle multi-path system model according to the order number and the network point number.
The vehicle information comprises the number of vehicles and/or the volume of the vehicles and/or the maximum volume of the vehicle container and/or the maximum load capacity of the vehicle container and/or the maximum driving range of the vehicle and/or the maximum operating time of the vehicle.
In the step c, a mixed path model is adopted, and the mathematical model established in the mixed path model is as follows:
wherein, defining the number of each mesh point as {1, 2, …, L };
the cost of distance transported from dot i to dot j is cij，
Each participating distribution vehicle is numbered K { K ═ 1,2, …, K };
the iteration upper limit value is Nm；
The defining variables are:
if the task of the network point i is completed by the delivery vehicle k, yik1, otherwise, yik＝0；
If the vehicle k runs from the node i to the node j, xijk1, otherwise, xijk＝0；
The hybrid path model comprises the maximum volume of vehicle packing and/or the maximum loading capacity of vehicle packing and/or the maximum driving mileage of vehicle running and/or the maximum working time of vehicle as constraint conditions, and the constraint model is as follows:
wherein the weight of the task goods of the network point i is gi；
The volume of the task goods of the network point i is qi；
Maximum payload of delivery vehicle k is Gk；
The maximum driving range of the distribution vehicle k is Sk；
Maximum volume of delivery vehicle k is Qk；
Total working time of delivery vehicle k is Tk；
The task goods working time of a distribution network point i of a distribution vehicle k is tki。
If the time window is opened, the time window of the hybrid path model comprises a first time window interval and a second time window interval, the time window interval from the mesh point i to the mesh point j is divided according to the working time, and the first time window interval is DBEi1—DBSi1The second time window interval is DBEi2—DBSi2And DBEi1＜DBSi1＜DBEi2＜DBSi2If the delivery vehicle arrives tki≤DBEi1Then t iski＝DBEi1(ii) a If the delivery vehicle arrives, DBSi1≤tki≤DBEi2Then t iski＝DBEi2(ii) a Other time periods arrive, tki＝tki。
The invention has the beneficial effects that:
the processing method for optimizing the vehicle path can select whether to open the time window and/or the line template according to the order number and the net point number, adopts the ant colony genetic fusion algorithm under the constraint of the time window, the line template and the vehicle information, optimizes the vehicle path after the input order number and the net point number, and obtains the optimal solution with lower cost.
Drawings
The following further describes embodiments of the present invention with reference to the drawings.
FIG. 1 is a flow chart of the process of the present invention.
FIG. 2 is a schematic diagram of a time window or route template for the single-vehicle multipath model of the present invention.
FIG. 3 is a schematic diagram of a time window or route template of the multi-vehicle multi-path system model of the present invention.
FIG. 4 is a schematic diagram of a time window or line template of the hybrid path model of the present invention.
Detailed Description
As shown in fig. 1 to 4, the processing method for optimizing the vehicle path in the present design can be applied to navigation of a concrete vehicle, a civil vehicle, and the like, and includes the following steps:
a. inputting order number, network point distance matrix and vehicle information;
b. setting whether a time window and/or a circuit template is opened or not;
c. and (3) taking the order number, the mesh point distance matrix, the vehicle information, whether a time window is opened or not and whether a line template is opened or not as constraint conditions, carrying out iterative computation by adopting an ant colony genetic fusion algorithm to obtain an optimal path, and updating mesh point pheromones.
And d, uploading the calculated result to a database of the cloud server.
In the preferred embodiment of the design, only the logistics path planning problem of a single starting point is considered, namely, the distribution vehicle starts from the starting point, passes through N network points and then returns to the starting point. This problem can be seen as a problem of queuing combinations of N mesh points, where the solution space is a factorial of N, and the optimal solution is the minimum of the sum of the distances of the N mesh point arrangements.
There are four cases according to the time window and whether the circuit template is opened: 1. a path model with a Time window opened and a Line Template opened, namely LT-CVRPTW (Line Template-Vehicle Routing protocol with Time Windows); 2. a path model with a Time window opened and a line template not opened, namely CVRPTW (vehicle Routing protocol with Time Windows); 3. a path model with a time window not opened and a Line Template opened, namely LT-CVRP (Line Template-Vehicle Routing protocol); 4. the path model with the time window not opened and the line template not opened, i.e. cvrp (vehicle Routing protocol).
The constraint conditions have the following points: path multi-time window opening or closing; opening or closing the circuit template; the vehicle information may include the number of vehicles and/or the volume of the vehicles and/or the maximum volume of the vehicle container and/or the maximum payload of the vehicle container and/or the maximum driving range over which the vehicle is operated and/or the maximum length of time over which the vehicle is operated.
The following limitations can be made in terms of functional requirements: the vehicle starts from the only distribution center and needs to return to the starting point after completing the delivery journey; the order is set with the latest arrival date, and the order of the latest day must be scheduled for delivery; each vehicle must set a line template number, and the number without opening the line template is set as-1; vehicles can only be delivered within a designated area.
In the step c, iterative computation is carried out within the range that the iteration upper limit value is not reached, and the method comprises the following steps:
c1, initializing ant colony and ant;
c2, selecting the next network point by the ant according to a deterministic exploration strategy or a stochastic exploration strategy;
c3, judging whether the ants go through all the network points, if yes, entering the step c4, and if not, returning to the step c 2;
c4, obtaining the circulating optimal ants after all the mesh points are processed;
c5, copying and crossing the cycle optimal ants and the global optimal ants obtained in the iterative process;
c6, performing exchange variation and inversion variation on globally optimal ants;
c7, judging whether the global optimal ants are subjected to iterative evolution, if so, entering a step c8, and otherwise, entering a step c 9;
c8, replacing the mesh point information of the original global optimal ant with the optimal solution obtained by iterative evolution, and entering step c 9;
c9, updating the site pheromone;
c10, judging whether the iteration number exceeds the iteration upper limit value, if so, outputting the result, and if not, returning to the step c 2.
Further, as shown in fig. 2 to 4, a single vehicle multipath model or a multi-vehicle multipath system model or a hybrid path model combining the single vehicle multipath model and the multi-vehicle multipath system model is established according to the order number and the network point number.
The use of a single-vehicle multipath model, which is similar to the traveler problem, TSP, minimizes the distance cost of this solution by finding a sequence of arriving nodes with known inter-node distance costs for a vehicle. The single-vehicle multi-path model needs to consider the maximum load capacity, the maximum driving mileage and the working time of the vehicle, and also considers the maintenance and the round trip of the vehicle, and the single-vehicle multi-path model can select four path models whether a time window and/or a route template is opened or not, complete all distribution tasks under the condition of only distributing one vehicle, and meet the corresponding requirements of customers on goods. Under the condition that a vehicle under a single-vehicle multi-path model must meet the condition that a line template is not opened, order tasks with small quantity can be independently completed, and the constraint requirements of a system can be met in a certain penalty mechanism, on the premise that the line template is opened, the line template numbers of the orders are the same, on the premise that the vehicle bound with the orders with the same line template number meets the constraint conditions, the distribution tasks of the orders can be independently completed, and the specific establishment of the single-vehicle multi-path model can be selected in the conventional technology.
The multi-vehicle multi-path system model is an extension of establishing on the basis of single-vehicle multi-path, when the number of orders is large, one vehicle is difficult to effectively complete the order distribution task, and at the moment, a plurality of vehicles are required to be arranged for path planning. The multi-vehicle multi-path model generally requires that a line template is closed to avoid the occurrence of distribution disorder accidents, the distributed vehicles must simultaneously meet the conditions of not exceeding the maximum load of the vehicles, the maximum driving mileage of the vehicles and the highest working time of the vehicles per day, then path planning is carried out under the constraint conditions of a time window and the line template, and the specific establishment of the multi-vehicle multi-path model can be selected in the conventional technology.
The optimal path model of the design is a mixed path model, a line template in the mixed path model must be opened, and the number of orders is large, so that a certain number requirement is met. Because the number of the distributed orders is large and the line template is opened, vehicles consistent with the line template participate in distributing network points matched with the line template number, and all models distributed by the line template are single-vehicle multi-path models; and some network points are not matched with vehicles, the network points of the unmatched vehicles finish distribution tasks by the vehicles without binding the line template numbers, because the network points are more in number, one vehicle is difficult to finish the current distribution task, a plurality of vehicles are required to participate in order distribution tasks, the distribution of non-line template network points forms a multi-vehicle multi-path model, and at the moment, the combination of the single-vehicle multi-path model and the multi-vehicle multi-path model forms a mixed path model.
The network points in the mixed path model can be divided into two parts, one part is the network points in the line template, the other part is the network points outside the line template, the vehicles complete the network point distribution task in the line template according to the number of the bound line template, the network points outside the distribution line template without binding the line template solve the optimal solution of the total paths and the lines of all the vehicles participating in distribution, and the combination of the single-vehicle multi-path model and the multi-vehicle multi-path model forms the mixed path model.
The mathematical model established in the hybrid path model is as follows:
wherein, defining the number of each mesh point as {1, 2, …, L };
the cost of distance transported from dot i to dot j is cijThe transport distance cost is mainly considered here as distance, assuming that the cost required for each distance is a fixed value.
Each participating distribution vehicle is numbered K { K ═ 1,2, …, K };
the iteration upper limit value is Nm；
The defining variables are:
if the task of the network point i is completed by the delivery vehicle k, yik1, otherwise, yik＝0；
If the vehicle k runs from the node i to the node j, xijk1, otherwise, xijk＝0；
The hybrid path model comprises the maximum volume of the vehicle container and/or the maximum loading capacity of the vehicle container and/or the maximum mileage of the vehicle operation and/or the maximum duration of the vehicle operation as constraint conditions, and the constraint model is as follows:
wherein the weight of the task goods of the network point i is gi；
The volume of the task goods of the network point i is qi；
Maximum payload of delivery vehicle k is Gk；
The maximum driving range of the distribution vehicle k is Sk；
Maximum volume of delivery vehicle k is Qk；
Total working time of delivery vehicle k is Tk；
The task goods working time of a distribution network point i of a distribution vehicle k is tki。
Here, the above formula (1) indicates that the total load of each vehicle cannot exceed the maximum load G of the vehicle; equation (2) indicates that the cargo volume carried by each vehicle cannot exceed the maximum volume Q of the vehicle; equation (3) represents the hypothetical T in this designkSetting the time to be 8 hours, indicating that the maximum working time of the vehicle cannot exceed 8 hours, and distributing the noon break time of the vehicle according to the actual situation to stagger the peak time of distribution; formula (4) indicates that the maximum travel distance of each vehicle cannot exceed the maximum travel distance S of the vehicle; meanwhile, the formula (5) also indicates that each vehicle is guaranteed to participate in the distribution task; and here, the circuit template must be opened, that is, LT ═ 1; the time window can be opened or closed, and TW is 1 or 0;
if the time window is opened, the time window of the hybrid path model comprises a first time window interval and a second time window interval, the time window interval from the mesh point i to the mesh point j is divided according to the working time, and the first time window interval is DBEi1—DBSi1The second time window interval is DBEi2—DBSi2And DBEi1＜DBSi1＜DBEi2＜DBSi2If the delivery vehicle arrives tki≤DBEi1Then t iski＝DBEi1(ii) a If the delivery vehicle arrives, DBSi1≤tki≤DBEi2Then t iski＝DBEi2(ii) a Other time periods arrive, tki＝tki。
DBE hereini1、DBSi1、DBEi2、DBSi2For example, two time window intervals are 8: 00-10: 00, 12: 00-14: 00; the working time value is divided into 0-2 and 4-6; facilitating machine calculations.
The design aims at the shortest path under the condition of single distribution point under the condition that the average cost of the lines is the same. Firstly, inputting each constraint condition, including vehicle information, whether a time window is opened or not, whether a circuit template is opened or not and the like, and recording a path after ants traverse all city network points and updating pheromones by utilizing an exploration strategy of an ant colony algorithm, which can be random exploration or deterministic exploration. The traditional ant colony algorithm is characterized in that pheromones are continuously superposed to find out the optimal solution under iteration, the algorithm is fused with a genetic algorithm, the path and the pheromone of the globally optimal ant are recorded, and the genetic algorithm is carried out with the path and the pheromone of the iteratively optimal ant to carry out cross and variation operations. Therefore, the global search capability of the genetic algorithm is fully utilized, the cross probability of the genetic algorithm is properly changed, the search range can be obviously improved, the discrete problem is better adapted, when the iteration upper limit is not reached, the steps are repeated, and finally, the ant path which tends to be stable is the output optimized path.
The design reasonably utilizes constraint conditions, the ant colony genetic fusion algorithm is copied and transformed after iterative computation, the shortest path is obtained, the performance index of the result is good, and the result can be uploaded to a server to be displayed.
The following are experimental results for proving that the performance indexes of the design result are better:
firstly, according to whether a time window and a circuit template are opened, four conditions are provided: the time window is not opened, the line template is not opened, the time window is not opened, the line template is opened, the time window is opened and the line template is opened, and three path models are proved according to the four experimental conditions. And simultaneously, the relations between the shortest distances of three orders with different quantities, namely order _25, order _50 and order _100, and the iteration times are respectively analyzed, and the specific contents of order _25, order _50 and order _100 are omitted here. Combining with the generally accepted standard for setting the parameters of the ant colony algorithm, taking values of three parameters of an information heuristic factor, an expected heuristic factor and an pheromone attenuation speed:
| parameter(s) | Of significance | Value taking |
| α | Information heuristic factor | 1.0 |
| β | Expectation heuristic factor | 4.0 |
| ρ | Rate of decay of pheromone | 0.75 |
(1) The bicycle multi-path model:
the experiment is carried out by adopting data in the order _25 file, the group of data comprises information of 25 nodes including a distribution center, because the nodes are not many, a single vehicle is used for distribution, the distribution vehicle returns to the distribution center after one round of distribution, and the path is a closed loop and indicates that one node of the vehicle only reaches once.
(2) Multi-vehicle multi-path model:
the experiment is carried out by adopting data in an order _50 file, the group of data comprises 50 network point information including a distribution center, a plurality of vehicles are distributed, a plurality of paths are arranged in the paths, and after the algorithm is adopted for planning, the vehicles with the same template number can only distribute orders to corresponding network points.
(3) Hybrid path model:
the data of the experiment is carried out by the data in the order _100 file, and the distribution center comprises 100 network points.
And finally comparing the shortest distances of the three orders of order _25, order _50 and order _100 with the relationship of two major constraints.
Therefore, when the order number is the same, the shortest distribution distance under the CVRP model and the LT-CVRP model is shorter, the shortest distribution distance under the corresponding CVRPTW model and the LT-CVRPTW model is longer, and the calculation result of the shortest distance is longer due to the opening of the explanation time window; the shortest distribution distance under the same CVRP and CVRPTW models is shorter, while the shortest distribution distance under the corresponding LT-CVRP and LT-CVRPTW models is longer, which shows that the calculation result of the shortest distance is increased due to the opening of the line template; in the same model, the larger the number of orders, the longer the calculated shortest distance. Under the premise of the same experimental conditions, the calculation results that the order number is increased, the line template is opened, and the time window is opened to be the shortest path are prolonged.
Compared with the conventional Ant colony algorithm, an Ant-cycle model and the like, the designed Ant colony genetic fusion algorithm performs 10 times of experiments on the Ant colony algorithm before improvement, the genetic algorithm and the improved fusion Ant colony and genetic algorithm respectively by using data provided by the order _25 file, and each parameter of the algorithm is set according to the general standard. The comprehensive performance of the algorithm can be compared according to the three indexes of the optimal performance index Eo, the time performance index ET and the robust performance index ER, and the smaller the values of the three indexes are, the better the comprehensive performance of the algorithm is represented.
The optimum performance index Eo, expressed as:
cband c represents a theoretical optimal value of the algorithm for solving the problem.
The time performance index ET is expressed by the formula:
Iarepresents the average value of iterations, T, when the termination condition is met after the algorithm is run for a number of times0Represents the average computation time, I, taken by the algorithm to iterate oncemaxIndicating the initially set maximum number of iterations.
Robust performance indicator ER, which is formulated as:
and ca represents the average value of the shortest paths obtained by the operation of the algorithm for multiple times, and c represents the theoretical optimal value of the algorithm for solving the problem.
The iteration number Imax set here is 500, the VRP model in this document is solved by using a separate ant colony algorithm, the pheromone updating strategy adopts the traditional optimization method of this iteration and the global optimization method, and c1 is set to 0 and c2 is set to 1.
The Ant-cycle model is also called an Ant-cycle model, a system of the model adopts a global optimization strategy to update pheromones, and the mode of updating the pheromones in the model only adopts the mutation operation of a genetic algorithm and only adopts cross mutation.
The Ant-Q system model algorithm pheromone is updated using a conventional global optimal update strategy and sets c1 to 0 and c2 to 1.
The improved algorithm pheromone updating strategy adopts a mixed strategy to update pheromones, data of an order _25 order is called to continuously test the algorithm for 10 times, and the values of c1 and c2 are set to be 0.5, so that the improved algorithm pheromone updating strategy becomes an improved global optimal updating strategy.
According to the calculation of performance indexes before and after improvement, three index results of an ant colony algorithm, an ant colony genetic algorithm before improvement and an ant colony genetic algorithm after improvement are obtained:
according to various indexes displayed by the table, on the premise that the iteration times Imax are the same, the ant colony algorithm is adopted to solve the VRP model of the graph independently, the duplication, crossing and variation operations of the genetic algorithm are avoided, and the ant colony algorithm has the advantage in time performance indexes. However, the ant colony algorithm does not find the optimal solution in the experiment, only finds the relatively optimal solution, and the ant colony algorithm does not take up the wind on the optimal performance index and the robust performance index, so the solution by adopting the ant colony genetic algorithm is better than that by adopting the ant colony algorithm.
Comparing the three performance index values of the Ant colony genetic algorithm of the Ant-cycle model with the three performance indexes of the improved Ant colony genetic algorithm, the three performance index values of the improved Ant colony genetic algorithm are smaller, and therefore the comprehensive performance of the improved Ant colony genetic algorithm is better.
Meanwhile, it can be seen that the Ant colony genetic algorithm of the Ant-Q system model and the improved Ant colony genetic algorithm can find the optimal solution in the experiment, so the optimal performance indexes are the same, but the time performance index of the improved algorithm is smaller than that before the improvement, the robustness is also smaller than that of the improved result, and therefore the comprehensive performance of the improved algorithm is better than that of the Ant colony genetic algorithm of the Ant-Q system model.
Further, the output of the design can show the scheduled path planning to a Baidu map through a cloud background web service such as Baidu map API and LBS cloud service, how to transmit the data to Baidu map API codes is also an important aspect, the LBS cloud related service can help solve the method, the LBS cloud can upload mass data of users to a cloud database of the users and can freely support large-flow access per second, and the Baidu map can display the data on the map in a pockmark form and can view the relevant information of the pockmark through clicking. But also refresh the latest website dynamics by constantly requesting.
The above description is only a preferred embodiment of the present invention, and the present invention is not limited to the above embodiment, and any technical means that can achieve the object of the present invention by basically the same means is within the scope of the present invention.
Claims (6)
1. A processing method for vehicle path optimization, comprising the steps of:
a. inputting order number, network point distance matrix and vehicle information;
b. setting whether a time window and/or a circuit template is opened or not;
c. taking the order number, the mesh point distance matrix, the vehicle information, whether a time window is opened or not and whether a line template is opened or not as constraint conditions, obtaining an optimal path by adopting ant colony genetic fusion algorithm iterative computation, and updating mesh point pheromones;
in the step c, a mixed path model is adopted, and the mathematical model established in the mixed path model is as follows:
wherein, defining the number of each mesh point as {1, 2, …, L };
the cost of distance transported from dot i to dot j is cijThe cost of the transport distance is mainly considered to be the distance, and the cost required by each distance is assumed to be a fixed value;
each participating distribution vehicle is numbered K { K ═ 1,2, …, K };
the iteration upper limit value is Nm；
The defining variables are:
if the task of the network point i is completed by the delivery vehicle k, yik1, otherwise, yik0; if the vehicle k runs from the node i to the node j, Xijk1, otherwise, xijk＝0；
The hybrid path model comprises the maximum volume of the vehicle container and/or the maximum loading capacity of the vehicle container and/or the maximum mileage of the vehicle operation and/or the maximum duration of the vehicle operation as constraint conditions, and the constraint model is as follows:
wherein the weight of the task goods of the network point i is gi；
The volume of the task goods of the network point i is qi；
Maximum payload of delivery vehicle k is Gk；
The maximum driving range of the distribution vehicle k is Sk；
Maximum volume of delivery vehicle k is Qk；
Total working time of delivery vehicle k is Tk；
The task goods working time of a distribution network point i of a distribution vehicle k is tki。
2. A processing method for vehicle path optimization according to claim 1, characterized in that: and d, uploading the calculated result to a database of the cloud server.
3. The vehicle path optimization processing method according to claim 1, wherein the iterative computation in step c is performed within the range that the iteration upper limit value is not reached, and the method comprises the following steps:
c1, initializing ant colony and ant;
c2, selecting the next network point by the ant according to a deterministic exploration strategy or a stochastic exploration strategy;
c3, judging whether the ants go through all the network points, if yes, entering the step c4, and if not, returning to the step c 2;
c4, obtaining the circulating optimal ants after all the mesh points are processed;
c5, copying and crossing the cycle optimal ants and the global optimal ants obtained in the iterative process;
c6, performing exchange variation and inversion variation on globally optimal ants;
c7, judging whether the global optimal ants are subjected to iterative evolution, if so, entering a step c8, and otherwise, entering a step c 9;
c8, replacing the mesh point information of the original global optimal ant with the optimal solution obtained by iterative evolution, and entering step c 9;
c9, updating the site pheromone;
c10, judging whether the iteration number exceeds the iteration upper limit value, if so, outputting the result, and if not, returning to the step c 2.
4. A processing method for vehicle path optimization according to claim 1, characterized in that: and c, establishing a single-vehicle multi-path model or a multi-vehicle multi-path system model or a mixed path model combining the single-vehicle multi-path model and the multi-vehicle multi-path system model according to the order number and the network point number.
5. The processing method for vehicle path optimization according to claim 4, wherein: the vehicle information comprises the number of vehicles and/or the volume of the vehicles and/or the maximum volume of the vehicle container and/or the maximum load capacity of the vehicle container and/or the maximum driving range of the vehicle and/or the maximum operating time of the vehicle.
6. The processing method of claim 5, wherein if the time window is open, the time window of the hybrid route model comprises a first time window interval and a second time window interval, the time window is divided from node i to node j according to the working time, and the first time window interval is DBEi1-DBSi1The second time window interval is DBEi2-DBSi2And DBEi1＜DBSi1＜DBEi2＜DBSi2If the delivery vehicle arrives tki≤DBEi1Then t iski＝DBEi1(ii) a If the delivery vehicle arrives, DBSi1≤tki≤DBEi2Then t iski＝DBEi2(ii) a Other time periods arrive, tki＝tki。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810845753.0A CN109117993B (en) | 2018-07-27 | 2018-07-27 | Processing method for optimizing vehicle path |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810845753.0A CN109117993B (en) | 2018-07-27 | 2018-07-27 | Processing method for optimizing vehicle path |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109117993A CN109117993A (en) | 2019-01-01 |
| CN109117993B true CN109117993B (en) | 2022-04-15 |
Family
ID=64863386
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810845753.0A Active CN109117993B (en) | 2018-07-27 | 2018-07-27 | Processing method for optimizing vehicle path |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109117993B (en) |
Families Citing this family (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109858703B (en) * | 2019-02-15 | 2021-04-27 | 中国环境监测总站 | Method and system for acquiring shortest path of sampling |
| CN112465180B (en) * | 2019-09-09 | 2023-09-26 | 北京京东振世信息技术有限公司 | Vehicle path planning method and device |
| CN110852524B (en) * | 2019-11-19 | 2023-04-18 | 腾讯科技（深圳）有限公司 | Method, device and equipment for determining time window and storage medium |
| CN111461402B (en) * | 2020-03-06 | 2024-03-26 | 上海汽车集团股份有限公司 | Logistics scheduling optimization method and device, computer-readable storage medium and terminal |
| CN111148110A (en) * | 2020-03-30 | 2020-05-12 | 汉腾汽车有限公司 | Vehicle networking road side unit deployment method based on genetic ant colony hybrid algorithm |
| CN111461624B (en) * | 2020-04-20 | 2023-07-11 | 上海东普信息科技有限公司 | Logistics line planning method, device, equipment and storage medium |
| CN111860957B (en) * | 2020-06-18 | 2023-12-29 | 浙江工业大学 | Multi-vehicle-type vehicle path planning method considering secondary distribution and balancing |
| CN111915078B (en) * | 2020-07-29 | 2023-09-26 | 中国烟草总公司湖北省公司 | Flexible cigarette distribution line planning method and system based on data driving |
| CN113177662A (en) * | 2021-05-08 | 2021-07-27 | 深圳小狮快送科技有限公司 | Method for optimizing distribution vehicle route with fixed distribution scheme constraint |
| CN113723675B (en) * | 2021-08-20 | 2024-03-26 | 深圳依时货拉拉科技有限公司 | Automatic scheduling method for spare parts and components and computer equipment |
| CN116795108B (en) * | 2023-06-09 | 2023-12-01 | 西南交通大学 | Intelligent unmanned vehicle distribution method based on multi-source sensing signals |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102508935A (en) * | 2011-09-22 | 2012-06-20 | 南京大学 | On-chip network mapping method based on ant-colony chaos genetic algorithm |
| CN104992242A (en) * | 2015-07-01 | 2015-10-21 | 广东工业大学 | Method for solving logistic transport vehicle routing problem with soft time windows |
| CN106251009A (en) * | 2016-07-27 | 2016-12-21 | 清华大学 | A kind of optimized algorithm of the Vehicle Routing Problems solving time window |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9536192B2 (en) * | 2014-06-23 | 2017-01-03 | International Business Machines Corporation | Solving vehicle routing problems using evolutionary computing techniques |
-
2018
- 2018-07-27 CN CN201810845753.0A patent/CN109117993B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102508935A (en) * | 2011-09-22 | 2012-06-20 | 南京大学 | On-chip network mapping method based on ant-colony chaos genetic algorithm |
| CN104992242A (en) * | 2015-07-01 | 2015-10-21 | 广东工业大学 | Method for solving logistic transport vehicle routing problem with soft time windows |
| CN106251009A (en) * | 2016-07-27 | 2016-12-21 | 清华大学 | A kind of optimized algorithm of the Vehicle Routing Problems solving time window |
Non-Patent Citations (1)
| Title |
|---|
| 《基于改进型遗传蚁群算法的TDOA多点定位研究》;唐菁敏 等;《通信技术》;20180710;第51卷(第7期);第1576-1578页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109117993A (en) | 2019-01-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109117993B (en) | Processing method for optimizing vehicle path | |
| Zhang et al. | Review of vehicle routing problems: Models, classification and solving algorithms | |
| CN103413209B (en) | Many client many warehouses logistics distribution routing resources | |
| CN110097234B (en) | Intelligent dispatching method and system for industrial cigarette transportation | |
| El-Sherbeny | Vehicle routing with time windows: An overview of exact, heuristic and metaheuristic methods | |
| Qin et al. | Reinforcement learning for ridesharing: An extended survey | |
| Jia et al. | A multi-objective optimisation algorithm for the hot rolling batch scheduling problem | |
| CN109002902A (en) | Subregion multistage fresh agricultural products dynamic vehicle method for optimizing route | |
| CN104700251A (en) | Maximum-minimum ant colony optimization method and maximum-minimum ant colony optimization system for solving vehicle scheduling problem | |
| CN103699982A (en) | Logistics distribution control method with soft time windows | |
| Yao et al. | A two-stage heuristic algorithm for the school bus routing problem with mixed load plan | |
| Li et al. | An ant colony optimization algorithm with adaptive greedy strategy to optimize path problems | |
| CN101814174A (en) | Optimization method of agricultural material chain operation logistics center site selection | |
| CN107358325A (en) | A kind of Location Selection of Logistics Distribution Center method, apparatus and computer-readable recording medium | |
| Shafahi et al. | SpeedRoute: Fast, efficient solutions for school bus routing problems | |
| Dubernet et al. | Including joint decision mechanisms in a multiagent transport simulation | |
| Asghari et al. | Environmental and social implications of incorporating carpooling service on a customized bus system | |
| Xu et al. | Dynamic vehicle routing using an improved variable neighborhood search algorithm | |
| Wang et al. | Cross-regional customized bus route planning considering staggered commuting during the COVID-19 | |
| Ma et al. | Balanced strategy based on environment and user benefit-oriented carpooling service mode for commuting trips | |
| El Bouzekri El Idrissi et al. | Evolutionary algorithm for the bi-objective green vehicle routing problem | |
| Zheng | Solving vehicle routing problem: A big data analytic approach | |
| CN110046851A (en) | Unmanned vehicle logistics method for allocating tasks based on Multi-Paxos | |
| Gan | A logistics distribution route optimization model based on hybrid intelligent algorithm and its application | |
| Ghandeharioun et al. | Real-time ridesharing operations for on-demand capacitated systems considering dynamic travel time information |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |